| 00:36 | <othermaciej_> | hmm I got a dozen new twitter followers all of a sudden |
| 00:42 | <Hixie> | you got a lot of press |
| 01:01 | <xtothey> | html5 discussion going on now at http://www.ustream.tv/channel/standardista. It's at the NY Web Standards Meetup. |
| 01:01 | <AryehGregor> | Requiring Flash rather than <video>, of course. |
| 01:03 | <xtothey> | Yeah, you missed a nice rant about Flash. |
| 01:04 | <othermaciej> | Hixie: I did? I didn't see any press other than some people retweeting the w3c announcement itself |
| 01:05 | <othermaciej> | oh crap I'm on Ars Technica |
| 01:05 | <Hixie> | yeah, my HTML5 google alert flagged all kinds of posts |
| 01:05 | <Hixie> | oddly i didn't see any headlines for paul |
| 01:05 | <Hixie> | (i didn't read any of the articles) |
| 01:18 | <othermaciej> | what does the word "boffin" mean? |
| 01:19 | <othermaciej> | one of these articles called me an "Apple WebKit boffin" and I am not sure whether I should be offended or flattered |
| 01:19 | <gavin_> | scientist/researcher, says http://www.answers.com/boffin |
| 01:19 | <gavin_> | http://www.urbandictionary.com/define.php?term=boffin |
| 01:20 | <gavin_> | flattered, I think :) |
| 01:21 | <Hixie> | othermaciej: flattered, it means a competent geek |
| 01:30 | <Hixie> | ok people |
| 01:30 | <Hixie> | stop filing bugs already |
| 01:30 | <Hixie> | the graph is going in the wrong direction |
| 01:36 | <othermaciej> | Hixie: you did ask people to file bugs... |
| 01:39 | <AryehGregor> | This 98.248.33.53 guy is a big fan of examples. |
| 01:43 | <Hixie> | yeah the examples were me |
| 01:43 | <Hixie> | :-) |
| 01:44 | <AryehGregor> | I might say something about revealing your home IP address, but I've assigned monoid.aryeh.name to point to my home IP address, so. I won't. |
| 01:44 | <AryehGregor> | (domain names make SSH to your home machine much more convenient!) |
| 01:46 | <Hixie> | my home machine that accepts incoming SSH is currently turned off waiting for snow leopard |
| 01:46 | <AryehGregor> | I never turn off my home desktop. I use it as an alarm clock, too. |
| 01:46 | <Hixie> | i have another home machine for that... |
| 01:46 | <AryehGregor> | It's really annoying when there's a hardware failure or something and it dies in the middle of the night. :( |
| 01:46 | <Hixie> | (i'm hoping snow leopard fixes the kernel panics i was getting running "isightcapture") |
| 01:46 | <AryehGregor> | Didn't I read something in your blog about how you use model trains for an alarm clock, or wanted to? |
| 01:47 | <Hixie> | i did, back in oslo |
| 01:47 | <Hixie> | that was quite effective |
| 01:47 | <Hixie> | my trains are all boxed up right now sadly |
| 01:47 | <AryehGregor> | That's classy. I just use: beep -f600 -l800 -r10 -D200 -n -f620 -l800 -r10 |
| 01:47 | <Hixie> | nowadays i use http://damowmow.com/playground/alarm.pl |
| 01:50 | <othermaciej> | I like to use "sleep 28800; open ~/Music/iTunes/iTunes\ Music/somerandom.mp3" |
| 01:50 | <AryehGregor> | Ugh, why is <select multiple> implemented so poorly by most UAs? Normal users can't figure out how to select multiple values at once. Ctrl-click is totally undiscoverable. |
| 01:51 | <AryehGregor> | They should do little checkboxes or something. |
| 01:51 | <othermaciej> | I think Mac users know how to use multi-select lists |
| 01:51 | <AryehGregor> | But Windows users don't, or you have no opinion on them? |
| 01:51 | <Hixie> | othermaciej: i used to do that (GNU sleep(1) is even better, it takes arguments like '8h'), but eventually i needed more, and that's how i ended up with that script (which does speech synthesis and all kinds of stuff) |
| 01:52 | <othermaciej> | the thing I really want from an alarm is 1-hour snooze button |
| 01:52 | <Hixie> | mine does that basically automatically |
| 01:52 | <othermaciej> | (although speech synthesis to remind me why I actually need to get up could be useful too) |
| 01:52 | <Hixie> | it says the time and a custom message, and repeats it at increasingly large intervals |
| 01:52 | <AryehGregor> | Man, I never use snooze buttons. I just get up. Maybe that's because I have something I do at 8 or 8:30 every morning, though, that I always have to be up for. |
| 01:53 | <Hixie> | 8:30am!!! |
| 01:53 | <Hixie> | jesus |
| 01:53 | <Hixie> | 4am - noon is bed time |
| 01:53 | <AryehGregor> | Sleeping in for me is waking up at 8 AM. :P |
| 01:54 | <othermaciej> | I like to plan to be able to wake up, then sleep another hour, then actually get up and do what I need to do |
| 01:54 | <othermaciej> | it's a lot easier for me to get up if I wake up twice |
| 01:55 | <AryehGregor> | It's morning prayers, you see. I have to say them every morning, and it has to be in the first half of the morning or so. Which means you have to start like 8:30 or 9:00 latest, at this latitude, if you want a consistent start time year-round. :) |
| 01:55 | <AryehGregor> | (and congregations usually keep a consistent start time year-round) |
| 01:55 | <AryehGregor> | othermaciej, why do you think Mac users know how to use multi-selects? |
| 01:56 | <othermaciej> | because list boxes are a common idiom, even in the file manager |
| 01:57 | <AryehGregor> | Yes, but are users forced to select multiple things? |
| 01:57 | <AryehGregor> | Ctrl-click works in Windows and Linux too in the file manager, but that doesn't mean everyone knows how to use it. |
| 01:57 | <othermaciej> | I think people know about shift-click and shift-downarrow and such |
| 01:57 | <othermaciej> | I could be wrong, it is just my subjective impression |
| 01:58 | <AryehGregor> | I suspect lots don't, but I tend to make a conscious effort to assume people are way less techy than I'd have thought, and I probably overshoot sometimes. |
| 06:44 | <Lachy> | Hixie, what's the point of the creationCallback in the openDatabase() method? |
| 06:45 | <Lachy> | I don't see any use for it since the database is still returned by the openDatabase() method. There doesn't appear to be any need to use a callback to get it. |
| 07:15 | <Mrmil> | so much noise about 'cite'... |
| 07:45 | <annevk2> | Lachy, it's called when you do this for the first time |
| 07:46 | <annevk2> | Lachy, so you can have a standalone function that is just for generating all the tables that are needed for your app |
| 07:46 | <annevk2> | Lachy, compare with the installation steps of e.g. WordPress |
| 08:01 | <eighty4> | gsnedders|work: pm? |
| 08:02 | <Lachy> | annevk2, perhaps I'm misunderstanding how it's supposed to work, but I think that would mean when you open the database, you have to wait for the callback to be invoked before doing anything with the database, but since it only gets invoked when the database is first created, how are you supposed to know when it's ready to be used? |
| 08:03 | <Lachy> | well, you get the database returned from the method, but before running any additional transactions, you still need to check whether or not it was created for the first time |
| 08:04 | <Lachy> | and if it was, you have to wait for the callback to be executed, where you can set up the tables required. |
| 08:06 | <Lachy> | basically, what I'm saying, is it sucks to have a semi-asynchronous/semi-synchrounous API like this with which you have no idea whether it will need to be run synchrounously or asynchronously when you open it :-( |
| 08:13 | <annevk2> | Lachy, no, that callback is invoked first |
| 08:13 | <annevk2> | Lachy, all other transactions will be queued behind it |
| 08:16 | <Lachy> | oh, right. I guess that's ok. It still sucks having so many damn callbacks in this api |
| 08:20 | <othermaciej> | hsivonen: are you around? |
| 08:22 | <hsivonen> | othermaciej: Yes |
| 08:22 | <othermaciej> | hsivonen: could you comment on the proposal from Jonathan Robie in <http://www.w3.org/Bugs/Public/show_bug.cgi?id=7059>;, in light of his clarifications? |
| 08:22 | <hsivonen> | othermaciej: OK. |
| 08:23 | <othermaciej> | hsivonen: the XQuery/XSL folks want to talk about this at TPAC, I am hoping we can get it resolved before then |
| 08:24 | <hsivonen> | othermaciej: I'm OK with the change if you guys change WebKit, too. :-) |
| 08:25 | <othermaciej> | hsivonen: I don't know the history of why we decided to do it this way |
| 08:25 | <othermaciej> | (i.e. whether actual sites broke or if it was just for logical reasons) |
| 08:25 | <hsivonen> | othermaciej: IIRC, ap said it was just in case. (Not his exact words.) |
| 08:25 | <othermaciej> | I was gonna say, he's the one who would know |
| 08:28 | <othermaciej> | hsivonen: I can't tell if Jonathan's proposed wording is for XPath or HTML5 |
| 08:28 | <annevk2> | what's he's proposing, how does that effect XSLT applied to HTML? |
| 08:28 | <othermaciej> | hsivonen: (or for whatever spec or specs define document.evaluate or XSLTProcessor |
| 08:28 | <annevk2> | s/effect/affect/ |
| 08:29 | <othermaciej> | annevk2: seems like no effect, since it's scoped to document.evaluate as written |
| 08:29 | <hsivonen> | annevk2: it doesn't affect XSLT applied to an HTML DOM compared to current HTML 5 spec text unless the tree also has no-namespace nodes introduced with createElementNS or by importing from an XML tree |
| 08:30 | <othermaciej> | hsivonen: did you find actual compat issues when you changed this in Gecko, or just broken test cases? |
| 08:30 | <hsivonen> | othermaciej: it needs to apply to *any* DOM API that causes XPath to be evaluated against an HTML DOM |
| 08:30 | <hsivonen> | othermaciej: including whatever the API is for invoking XSLT from JS |
| 08:31 | <hsivonen> | othermaciej: making no-namespace name expressions match http://www.w3.org/1999/xhtml nodes in HTML trees is needed for compat. |
| 08:31 | <hsivonen> | othermaciej: making no-namespace name expressions also match no-namespace element nodes is in Gecko solely for uniformness with the WebKit behavior |
| 08:32 | <annevk2> | sorry, phrased my question wrong, I was wondering whether that'd be a problem |
| 08:32 | <othermaciej> | hsivonen: I see, so the WebKit change you are talking about is to *not* match no-namespace element nodes in HTML trees? |
| 08:32 | <othermaciej> | hsivonen: I don't see a problem with doing that |
| 08:32 | <othermaciej> | hsivonen: I misunderstood what you meant at first |
| 08:33 | <hsivonen> | othermaciej: yes |
| 08:34 | <othermaciej> | we'd just have to tweak our isHTMLElement/isHTMLDocument special case slightly |
| 08:38 | <hsivonen> | othermaciej: commented |
| 08:38 | <hsivonen> | on the bug that is |
| 08:38 | <othermaciej> | hsivonen: thanks! |
| 08:59 | <hsivonen> | I like the new meta@scheme example |
| 09:02 | <Lachy> | hsivonen, which example are you referring to? |
| 09:03 | <hsivonen> | Lachy: the one with ORLY |
| 09:30 | <hsivonen> | tantek: does gmpg.org get a lot of automated profile fetches? if so, do they come from apps other than GRDDL tools? |
| 09:38 | Philip` | wonders why a fractal is more infinitely detailed than e.g. the edge of a circle, which you can zoom in on forever and it will constantly look different |
| 09:39 | <poe> | it isn't, it's just more suprising |
| 09:41 | <othermaciej> | there's a well-defined way in which you can say it's "more detailed" |
| 09:41 | <othermaciej> | the edge of a circle is smooth, in a mathematically definable way |
| 09:41 | <othermaciej> | it's continuously differentiable |
| 09:41 | <othermaciej> | most fractals are nowhere differentiable - they are perfectly jagged |
| 09:50 | <jgraham> | gsnedders|work: Trying to introduce paradoxes into the spec I see |
| 09:50 | <tantek> | hsivonen - good q - I'll have to check the logs |
| 09:50 | <gsnedders|work> | jgraham: Indeed. I mean, the CSS WG has already done it with CSS 3 Selectors. |
| 09:50 | <gsnedders|work> | jgraham: We can't let the CSS WG out do us on paradoxes! |
| 09:51 | <tantek> | BTW FWIW I agree with having an explicit "element that goes between <header> and <footer>" |
| 09:51 | <tantek> | AKA <content> |
| 09:51 | <tantek> | people *will* make one up or (ab)use article/section |
| 09:51 | <tantek> | annevk3 is right |
| 09:52 | <tantek> | gsnedders - what paradox do you know of in the Selectors spec? |
| 09:52 | <jgraham> | We should make a RFC that lists all RFCs that don't refer to themselves |
| 09:53 | <gsnedders|work> | tantek: See ::selection |
| 09:53 | <tantek> | yeah - that was always a tough one |
| 09:53 | <gsnedders|work> | "This section intentionally left blank. (This section previously defined a ::selection pseudo-element.)" |
| 09:53 | <tantek> | thought we figured that one out though with a combination of pseudo-pseudo-elements and limited properties |
| 09:53 | <gsnedders|work> | (The paradox is the fact it says it is blank, hence making it not blank.) |
| 09:54 | <tantek> | "this is not a blank page" |
| 09:54 | <tantek> | nor a pipe for that matter |
| 09:54 | <othermaciej> | Russel's Transfer Protocol |
| 09:54 | <gsnedders|work> | (I just proposed on public-html to have explicitly blank sections in authoring view of the spec to keep section numbers the same, which is why jgraham brought it up) |
| 09:55 | <tantek> | gsnedders - what a waste of paper! ;) |
| 09:55 | <gsnedders|work> | tantek: Who prints specs? :P |
| 09:56 | <tantek> | people that work at companies that make printers, and/or print user agents? |
| 09:57 | <gsnedders|work> | tantek: Yes, I know some actually do. But I think the majority of people who do will be printing out the full spec, not the authoring view, so there wouldn't be the "blank" sections. |
| 10:00 | <Philip`> | In case people do actually print it, the spec ought to be padded with lots of blank pages to save ink |
| 10:03 | <Lachy> | then people should print with a really really tiny font size to reduce the total number of pages needed, counteracting the excess gained from the blank page padding |
| 10:04 | <beowulf> | people who want signed copies when they attend the $1000/hr HTML5 course |
| 10:05 | <gsnedders|work> | No, it was $1500/hr! Get it right! |
| 10:05 | <beowulf> | pfft |
| 10:11 | <jgraham> | I thought it was $1500/session. Were the sessions only 1hr? |
| 10:11 | <Lachy> | anyone remember the link for those courses? |
| 10:11 | <jgraham> | http://skillsmatter.com/course/ajax-ria/enterprise-comet-ajax |
| 10:11 | <jgraham> | (they are 3 days which is more like $500/day or $80 per person per hour) |
| 10:11 | <beowulf> | whatever, you'll surely recoup all costs with a signed copy of the html5 spec |
| 10:11 | <beowulf> | hixie, shelly, the adobe dudes, that apache guy... |
| 10:11 | <beowulf> | priceless |
| 10:11 | <beowulf> | antiques roadshow 2050, here i come |
| 10:11 | <Lachy> | beowulf, would it still be priceless if Hixie just digitally signed it, and could then sell unlimited copies without wasting paper? |
| 10:11 | <beowulf> | :) |
| 10:14 | <jgraham> | It's only worth anything if everyone in the acknowledgements section signs it |
| 10:15 | <Lachy> | jgraham, actually, it's worth whatever the market will pay for it, regardless of how many people have signed it |
| 10:15 | <gsnedders|work> | Hmm, that sounds like a challenge. |
| 10:15 | <Lachy> | it'd be interesting to put this to the test. Several of us print out and sign our own copy, and see which one fetches the most on ebay |
| 10:16 | <gsnedders|work> | I still want to try and get everyone named in the spec to sign it. |
| 10:16 | <gsnedders|work> | I guess that means getting Hendrel and Pillar to sign it |
| 10:17 | <Lachy> | gsnedders|work, do you still have that TPAC cabal-attendance sheet from last year? How much do you reckon that would fetch? |
| 10:17 | <gsnedders|work> | Lachy: It's at my parent's home somewhere, I think I know where. Dunno. |
| 11:02 | <beowulf> | would <body><main> ... </main></body> be abuse of <main>? |
| 11:02 | <beowulf> | because that's what i think a lot of people would use it for |
| 11:03 | <annevk2> | yes |
| 11:03 | <annevk2> | that's easy to flag though |
| 11:04 | <beowulf> | cool |
| 11:34 | <annevk3> | quite often when I see boolean attributes in use people just stick true in there... |
| 11:34 | <annevk3> | e.g. loop="true" |
| 11:35 | <takkaria> | XML really shouldn't have got rid of boolean attributes |
| 11:49 | <hsivonen> | takkaria: SGML didn't have boolean attributes, so one might argue XML never got rid of them |
| 11:50 | <hsivonen> | takkaria: I think it's good XML got rid of what SGML had here |
| 11:52 | <takkaria> | oh, I thought it did |
| 11:52 | <takkaria> | anyway, boolean attributes would have been a good idea |
| 11:54 | <hsivonen> | takkaria: SGML had something crazier: if the DTD had an enumeration of possible attribute values, specifying just the value implied the attribute name |
| 11:55 | <hsivonen> | takkaria: so in SGML, <input type checked> makes checked the value of the attribute that implies the attribute name checked |
| 11:55 | <hsivonen> | so in theory <input checkbox> should imply <input type=checkbox> |
| 11:56 | <hsivonen> | in SGML |
| 12:02 | <annevk3> | hsivonen, or <input password> implied <input type=password> |
| 12:02 | <annevk3> | oh wait, you just mentioned that one :) |
| 12:02 | annevk3 | confused checkbox with checked |
| 12:28 | <zcorpan> | history navigation and media elements is interesting |
| 12:43 | <takkaria> | hsivonen: mildly terrifying, that |
| 12:44 | <annevk3> | zcorpan, "interesting", I suppose :) |
| 12:46 | <zcorpan> | annevk3: yes |
| 12:47 | <zcorpan> | as in causing crashes if you're not careful and requiring lots of work to make events not fire |
| 13:17 | <Lachy> | annevk, how would <body><main>...</main></body> be an abuse? |
| 13:18 | <Lachy> | maybe if the author did <body><main><header/>...<footer/></main></body> instead of <body><header/><main>...</main><footer/></body> |
| 13:18 | <Lachy> | (though we really should use <content> instead of <main>, if at all) |
| 13:23 | <beowulf> | Lachy: my question was about <body><main><header/>...<footer/></main></body>, fwiw |
| 13:24 | <Lachy> | beowulf, ah, ok. In that case, yes, I agree with annevk2 |
| 13:26 | <beowulf> | http://www.stuffandnonsense.co.uk/archives/naming_conventions_table.html # in this, as many have main as inside the body as have it inside <header/> <footer/> |
| 13:26 | <beowulf> | but if like annevk3 says it's easily flagged, no worries |
| 13:27 | <zcorpan> | "must not have header or footer descendants without a sectioning element in between" |
| 13:43 | <beowulf> | should the div section in the spec not mention it's use in styling, it does in the example text I suppose... |
| 13:52 | <gsnedders|work> | WebKit fun: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/221 |
| 13:54 | <gsnedders|work> | (Any DOM manipulation doesn't get rendered, too) |
| 13:54 | <poe> | (and safari crashed.) |
| 13:55 | <gsnedders|work> | If you make the script not run you can modify it fine |
| 14:15 | <eighty4> | gsnedders|work: pm? |
| 14:15 | <TabAtkins> | othermaciej: The average person does *not* know about shift+click or ctrl+click, and *especially* not shift+down arrow. Keycombos are voodoo magic to the average person. I was tech support; I know. |
| 14:16 | <hsivonen> | TabAtkins: anyone who has used Mac vector drawing programs any time since the 1980s knows about shift-click and command-click |
| 14:17 | <TabAtkins> | I can assure you that Windows users do not know about this. |
| 14:17 | <AryehGregor> | How many people (Mac users or otherwise) have used Mac vector drawing programs? |
| 14:17 | <TabAtkins> | And I'm not convinced that knowledge of keycombos for specific programs translates into knowledge of key combos in general. |
| 14:17 | <TabAtkins> | Especially when they're completely different. Context is lost on the average user. |
| 14:18 | <hsivonen> | TabAtkins: I'm ready to believe Windows users don't know about multiple selection, but it should be pretty common knowledge among Mac users |
| 14:18 | <TabAtkins> | One of you mentioned that it was used often in the file manager or whatnot? |
| 14:18 | <AryehGregor> | TabAtkins, you know, I think we need :changed/:unchanged. Then you could do :invalid:changed:after { contents: " !"; color: red; font-weight: bold; } or such. (Should :changed apply if the UA prefills the value?) |
| 14:19 | <TabAtkins> | How is it done so that it differs from the windows file system? |
| 14:19 | <TabAtkins> | (haven't has experience with Macs in 10+ years) |
| 14:19 | hsivonen | imagines anyone who doesn't know about shift/command-click is in the world of pain with the Finder |
| 14:19 | <AryehGregor> | TabAtkins, Macs have a sort of list-box file manager. |
| 14:19 | <TabAtkins> | AryehGregor: Hrm, yes. Let's work out details. One sec, though - I just arrived at work, and need to go put on clothes. |
| 14:19 | <AryehGregor> | TabAtkins, http://www.qdev.de/mac/pics/exifrenamer_folder_after_backdrop.png |
| 14:20 | <hsivonen> | TabAtkins: shift-click selects the range between current selection and the click |
| 14:20 | <hsivonen> | TabAtkins: command-click adds the clicked item to selection |
| 14:20 | <AryehGregor> | hsivonen, the same is true on Windows, but I bet most Windows users don't know about it. |
| 14:20 | <AryehGregor> | How often do users really *need* to select multiple files? E.g., to delete a bunch, they can just delete them one by one. |
| 14:20 | <AryehGregor> | People do stuff like that. |
| 14:21 | <hsivonen> | AryehGregor: it's generally painful to watch Windows users struggle with their UI |
| 14:21 | <AryehGregor> | Hmm, on Windows they can also select multiple files by click and drag a box, though. |
| 14:21 | <hsivonen> | it seems to me that when they are frustrated enough they stop learning |
| 14:21 | <hsivonen> | but Mac users don't get demotivated and learn these shortcuts |
| 14:21 | <AryehGregor> | It seems to me that most people just don't learn UI unless it's explicitly taught to them. |
| 14:22 | <AryehGregor> | That kind of person is probably less likely to use a Mac, though. |
| 14:22 | <AryehGregor> | (much as I suspect the large majority of Linux users know about Shift-/Ctrl-click) |
| 14:22 | <zcorpan> | click-drag should work with <select multiple> |
| 14:22 | <AryehGregor> | The least competent computer users usually wind up with Windows because they don't really know what an operating system is anyway and just get whatever the store sells, or whatever. |
| 14:23 | <TabAtkins> | hsivonen: I think your experiences with Mac users may be somewhat dated. IMXP modern mac users have regressed to the mean. |
| 14:23 | <hsivonen> | could be |
| 14:23 | <AryehGregor> | Click-drag doesn't seem to work with <select multiple> on Chrome/Linux |
| 14:23 | <AryehGregor> | . |
| 14:24 | <TabAtkins> | AryehGregor: I'm not sure how that screenshot displays anything different from Windows. Am I missing something? |
| 14:24 | <AryehGregor> | TabAtkins, well, it's clearly different from the default Windows layout, but I'm not sure the difference is relevant. |
| 14:24 | <TabAtkins> | hsivonen: Tragedy of the commons, I guess. ^_^ |
| 14:24 | <AryehGregor> | . . . Tragedy of the commons is something totally unrelated. |
| 14:24 | <TabAtkins> | AryehGregor: Yes, slightly. |
| 14:25 | <TabAtkins> | That would be the use of the ^_^. I was being ironic in misusing the word to indicate that the common people have infiltrated the Mac. |
| 14:25 | <AryehGregor> | Tragedy of the commons is if Mac users use up all the Shift-clicking in the world so that they can't Shift-click anymore after ten years. |
| 14:26 | <TabAtkins> | Like Lispers using up all the parens? |
| 14:26 | <TabAtkins> | Or perlers using up all the linenoise, so TVs won't be able to display static anymore? |
| 14:27 | <TabAtkins> | So, anyway, :changed. I like the concept. |
| 14:27 | <jgraham> | I would be astonished if the majority of users knew about these shortcuts, Macs or not. I would be astonished if there was ever a time since the Mac went mass-market that the majority of mac users knew about them |
| 14:28 | <AryehGregor> | Me too. |
| 14:28 | <AryehGregor> | It's way too easy to assume most people are at least slightly techy like us. |
| 14:28 | <jgraham> | Frankly I have difficulty remembering which key does which thing |
| 14:28 | <TabAtkins> | Doesn't match when you first load the page, even if @value is specified. Does match as soon as the user causes a change event (or UA equivalent). |
| 14:28 | <TabAtkins> | Only ambiguity (maybe) is what to do if script alters the value. |
| 14:28 | <AryehGregor> | I've read enough tech support horror stories to not be so optimistic. |
| 14:28 | <AryehGregor> | TabAtkins, or what if the UA prefills it? |
| 14:28 | <TabAtkins> | In what manner? |
| 14:29 | <AryehGregor> | Like if you go to a login page and it prefills your username. |
| 14:29 | <TabAtkins> | I feel like I want to classify that with scripts filling it. |
| 14:30 | <TabAtkins> | Though it may be okay to let the UA decide. |
| 14:31 | <TabAtkins> | Depending on what they think it's closer to - them just rendering a static @value, or them typing for the user. |
| 14:32 | <TabAtkins> | I think scripts should probably trigger :changed just like a user filling values in. |
| 14:34 | <AryehGregor> | Yes, I'd tend to agree. |
| 14:34 | <AryehGregor> | If they don't want to trigger it, they can change the default value instead of the current value. |
| 14:35 | <AryehGregor> | What if the user changes it, then changes it again so it's equal to the default value? |
| 14:35 | <AryehGregor> | Maybe even without committing the first change? |
| 14:36 | <AryehGregor> | I would say probably don't trigger it. |
| 14:36 | <AryehGregor> | (I mean, it should be :unchanged) |
| 14:37 | <TabAtkins> | Don't trigger it - it wouldn't fire a change event in that case. |
| 14:37 | <AryehGregor> | But what if they changed it, committed the change, and went back to change it back? |
| 14:37 | <TabAtkins> | Then it still matches :changed. |
| 14:37 | <AryehGregor> | Yes, I agree. |
| 14:38 | <TabAtkins> | Yup, the user has meaningfully interacted with it, so we're good. |
| 14:38 | <AryehGregor> | :changed would match HTML's idea of dirtiness, I think. |
| 14:38 | <hsivonen> | Hixie: the spec should use @font-face to supply proper fonts to all viewers |
| 14:38 | <TabAtkins> | How, hm, how does one change the default value of an input without changing the current value? This may be my jQuery thinking collapsing ideas in my head. |
| 14:39 | <Lachy> | I'm surprised that windows users struggle with the concept of shift-/control-clicking for selection |
| 14:39 | <jgraham> | s/windows// |
| 14:39 | <AryehGregor> | TabAtkins, well, you could just change the "value" attribute . . . |
| 14:39 | <AryehGregor> | Lachy, do you have evidence that *average* Mac users don't? |
| 14:39 | jgraham | is a Mac/Linux user and is tired of Mac-elitism on this topic |
| 14:39 | <TabAtkins> | Yeah, then it's me collapsing ideas. ^_^ I tend to think of .val() and .attr("val",) the same. |
| 14:39 | <Lachy> | no, I'm not commenting on mac users |
| 14:40 | <TabAtkins> | Bleh, .attr("value",) |
| 14:40 | <Lachy> | given that windows explorer has the behaviour, it seems like it should be fairly common knowledge |
| 14:40 | <Lachy> | and it always just seemed intuitive to me |
| 14:40 | <TabAtkins> | Lachy: It's simply not *necessary* behavior for most people. |
| 14:40 | <TabAtkins> | And a goodly fraction understand the "drag a box" behavior, which is a good substitute. |
| 14:41 | <AryehGregor> | Lachy, it's totally arbitrary, though, if you think about it. It's not intuitive at all, or discoverable. The only possible way you could figure it out is either by reading about it, being told about it, or banging your keyboard randomly while selecting stuff. |
| 14:41 | <TabAtkins> | Also: the normal display of Windows folders is "Thumbnail" rather than "List". |
| 14:41 | <AryehGregor> | The average computer user doesn't read about computers or talk to people about them, unless they have a problem they need to solve. |
| 14:41 | <jgraham> | The drag a box behaviour is more discoverable because it's something that novice users often do by accident |
| 14:41 | <Lachy> | although, IIRC, MS Office drawing functions (in Word, Powerpoint, etc.) did mess it up, and made Shift+click behave like ctrl+click should |
| 14:41 | <TabAtkins> | jgraham, yup. |
| 14:42 | <jgraham> | whereas there's no obvious reason that ctrl+click would do something different to z+click or f1+click |
| 14:42 | <AryehGregor> | I wouldn't go that far. Ctrl is a modifier key, I think most users see a difference between it and Z or F1. |
| 14:42 | <Lachy> | I don't know how I learned about it, but I definitely didn't read about it. I guess I must have been shown at some point or seen someone else do it |
| 14:42 | <AryehGregor> | But it's not obvious that it would do anything special in the magical narrow case where you already have something selected and are clicking on something else of the same type in a case where you can select both of them. |
| 14:43 | <AryehGregor> | Lachy, yeah, me too. I assume at some point I saw someone do it. |
| 14:43 | <TabAtkins> | AryehGregor: I think the primary difference people see is that Ctrl doesn't do anything most of the time. ^_^ |
| 14:43 | <AryehGregor> | I know that's how I figured out that you can expand and collapse tree views with the arrow keys. |
| 14:43 | eighty4 | knows that ~50% of the mac users dosesn't know of two fingers scroll |
| 14:43 | <TabAtkins> | I'm pretty sure I learned about it from a "Windows 95 for dummies" book when I was 8 or someething. |
| 14:43 | <AryehGregor> | TabAtkins, hmm, true. Shift is maybe a better example. |
| 14:43 | <AryehGregor> | TabAtkins, was Windows 95 released when you were 8? |
| 14:43 | <TabAtkins> | Well, I suppose I had to be at least 10 for that. |
| 14:44 | <Lachy> | also, I remember lots of sites that used to use <select multiple> used to say things like "Hold Ctrl to select multiple items" or similar |
| 14:44 | <Lachy> | though the use of <select multiple> seems to have declined in favour of checkboxes |
| 14:44 | <AryehGregor> | Oh, yeah, that's true. |
| 14:44 | <AryehGregor> | MediaWiki used to do that in some places. Then we switched to checkboxes. :P |
| 14:45 | <TabAtkins> | <select multiple> is just *way* too user-unfriendly. The fact that you *have* to be holding Ctrl to multi-select, and that it wipes out your selection if you forget at some point, is just evil. |
| 14:45 | <TabAtkins> | I've never used it in a design, and I don't intend to. |
| 14:46 | <Lachy> | it would be nice if the multi-select widget had check boxes of some kind beside each item in the list, that just required clicking without a modifier key |
| 14:46 | <TabAtkins> | Yeah. |
| 14:46 | <Lachy> | like how many installation wizards do it in the custom install dialogs |
| 14:46 | <AryehGregor> | That's what I think. |
| 14:46 | <poe> | this is a pretty nice layer ontop of select multiple http://code.google.com/p/jquery-asmselect/ |
| 14:46 | <AryehGregor> | There's no reason everyone has to use this cruddy UI. |
| 14:48 | <TabAtkins> | Yeah, I really like that redesign of <select>. It works well, and is intuitive. |
| 14:49 | <TabAtkins> | Of course, just using a bit list of checkboxes within a container with max-height and overflow:scroll set works pretty equivalently. |
| 14:54 | <TabAtkins> | Aryeh: You wanna write up :changed/:unchanged on the list? |
| 14:57 | <mpt> | TabAtkins, browsers could make <select multiple> easy to use very simply: just add a checkbox to each item. |
| 14:57 | <TabAtkins> | mpt: I agree. But no one has. ;_; |
| 14:57 | <mpt> | TabAtkins, and I think somewhere in bugzilla.mozilla.org there's a bug report from me to that effect. :-) |
| 14:58 | <TabAtkins> | And until at least IE and FF do it, I'm not willing to use <select multiple> |
| 14:58 | mpt | realizes Lachy just suggested the same thing |
| 14:59 | <TabAtkins> | Nod. It's silly how basic the solution is, and that it still hasn't been put into practice. |
| 14:59 | <mpt> | What does <select multiple> look like on an iPhone? |
| 15:00 | <TabAtkins> | Is a good question. |
| 15:00 | <mpt> | (since <select> looks very different) |
| 15:00 | <annevk3> | does Ubuntu have them mpt? :) |
| 15:00 | <annevk3> | btw, the wifi donation thing worked out |
| 15:00 | <TabAtkins> | Hrm... |
| 15:00 | <mpt> | annevk3, no Web browsers currently shipping with Ubuntu do, if that's what you mean |
| 15:00 | TabAtkins | goes and checks what it looks like on his phone with Opera Mini. |
| 15:01 | <annevk3> | i meant the OS |
| 15:01 | <mpt> | annevk3, have what, specifically? Lists where you can select more than one item with the Ctrl key? |
| 15:01 | <mpt> | Or lists with checkbox items? |
| 15:01 | jgraham | assumes the latter |
| 15:02 | <annevk3> | mpt, presumably ubuntu has multiple select as widget; i wondered whether you had the impl fixed there |
| 15:03 | <TabAtkins> | Ah, good job Opera Mini team. |
| 15:03 | <mpt> | That's up to the application. I don't recall any current application that expects people to know Ctrl click, apart from the file manager. |
| 15:03 | <TabAtkins> | A <select multiple> opens to a new screen filled with checkboxes. |
| 15:03 | <jgraham> | I think it is more accurate to say that Gtk has multiple-select widgets and that they din't have checkboxes |
| 15:03 | <jgraham> | *don't |
| 15:04 | <jgraham> | But I don't recall actually seing a multiple-select widget |
| 15:04 | <TabAtkins> | annevk3: Pass along my thumbs up to whoever does that. |
| 15:05 | <jgraham> | TabAtkins: takkaria is more likely to know I guess |
| 15:05 | <TabAtkins> | takkaria is Opera dev? |
| 15:06 | <TabAtkins> | I really only know for sure like two people who are on Opera - Anne and Hakon. |
| 15:06 | <TabAtkins> | I know there are lots others who hang around, but I don't know who they are. |
| 15:06 | <TabAtkins> | I guess Lachy? |
| 15:07 | <Lachy> | no, takkaria doesn't work for Opera |
| 15:07 | <jgraham> | Yep. And me and gsnedders|work for a year and takkaria for the summer and lots of other people who are on the list but not IRC |
| 15:07 | <TabAtkins> | kk |
| 15:08 | <Lachy> | jgraham, is takkaria a summer intern in sweden? |
| 15:08 | <jgraham> | takkaria is on the mini team for the summer |
| 15:08 | <jgraham> | in Sweden, yes |
| 15:08 | <gsnedders|work> | (That's me for a year, not jgraham also.) |
| 15:08 | <TabAtkins> | takkaria: Pass along thumbs up for Mini's impl of <select multiple>. |
| 15:08 | <jgraham> | I guess commas would have helped |
| 15:08 | <gsnedders|work> | (Though technically from tomorrow until the 4th October I'm not an Opera guy) |
| 15:08 | <TabAtkins> | Nod. I thought you were calling yourself an intern, jgraham. ^_^ |
| 15:09 | <jgraham> | No I'm trapped here |
| 15:09 | <jgraham> | ;) |
| 15:09 | <Lachy> | gsnedders|work, I thought you were only going to be with us for a few weeks |
| 15:09 | gsnedders|work | notes there's nobody standing in the doorway |
| 15:09 | gsnedders|work | wonders what traps jgraham |
| 15:09 | <gsnedders|work> | (I'm in the same room as jgraham) |
| 15:10 | <gsnedders|work> | TabAtkins: zcorpan is Opera as well (and is really Simon Pieters) |
| 15:10 | <TabAtkins> | Ah, I never connected the two names. |
| 15:10 | <jgraham> | Oh yeak, sorry I forgot |
| 15:10 | <jgraham> | *yeah |
| 15:10 | <Lachy> | gsnedders|work, jgraham has a really strong electromagnet in his chair and metal implants in his legs that keep him trapped |
| 15:11 | <gsnedders|work> | Nah, doesn't look like it |
| 15:11 | <jgraham> | ^this is true |
| 15:11 | <gsnedders|work> | Unless it's under the floor… |
| 15:13 | <TabAtkins> | You know, I've got a ton of emails on hand in a database of mine too. |
| 15:13 | <TabAtkins> | Might be fun to run AryehGregor's query against them too |
| 15:50 | <TabAtkins> | Well, I've got 441 or so distinct email addresses harvested from our customers, and they're all valid according <input type=email> validation. |
| 15:51 | TabAtkins | will see if he can get access to our full email lists - these are just the emails given to one particular app he wrote. |
| 16:39 | <gsnedders|work> | heycam: yt? |
| 16:41 | <gsnedders|work> | heycam: ignore me |
| 16:46 | <takkaria> | TabAtkins: yeah, it's quite nice, isn't it? though we got a bug report complaining about it a couple weeks back |
| 16:46 | <TabAtkins> | takkaria: Dude, how can you complain about it? |
| 16:46 | <takkaria> | we gave the response "it works like this on purpose" :) |
| 16:46 | <TabAtkins> | Heh. |
| 16:51 | TabAtkins | notes that Century Gothic is a *really nice* font. |
| 16:51 | jgraham | grumbles |
| 16:52 | <jgraham> | Hey, Century Gothic. That is a cheering thought |
| 16:52 | <gsnedders|work> | Says the grumpy old man |
| 16:52 | <TabAtkins> | ...I'm not quite sure how to respond to that. |
| 16:53 | <jgraham> | TabAtkins: What the idea that nice typography cazn be cheering? |
| 16:53 | <jgraham> | Or that I can be grumpy? |
| 16:53 | <TabAtkins> | jgraham: I wasn't sure how much sarcasm was in there. ^_^ |
| 16:53 | <TabAtkins> | I'm just happy that Century Gothic scales so well on-screen. |
| 16:54 | <TabAtkins> | Very readable down to 10px size. |
| 16:55 | <TabAtkins> | Dammit, FF3.5 doesn't support calc() yet. >_< |
| 20:10 | <atwilson> | Hixie: Not sure why you find it valuable to allow sending "lame duck" ports via postMessage()? This seems like it's always an error and should be reflected as an exception. |
| 20:11 | <atwilson> | Why would we want to allow this? |
| 20:13 | <Hixie> | isn't that the whole point of that thread? |
| 20:14 | <atwilson> | Hixie: no. That thread is about *closing* ports, not sending *cloned* ports. |
| 20:14 | <atwilson> | A closed port != a cloned port. |
| 20:15 | <atwilson> | In particular: |
| 20:15 | <atwilson> | If I have two entangled ports, port1 and port2, and I do "port1.postmessage(); port1.postMessage(); port1.close(); |
| 20:16 | <atwilson> | Someone should still be able to send port1/port2 around through postmessage, because there could still be queued up but undelivered messages associated with those ports. |
| 20:16 | <atwilson> | But once I've done this: |
| 20:16 | <atwilson> | port1.postmessage("msg", port2) |
| 20:17 | <atwilson> | Then there's no reason to allow doing port1.postMessage('msg2', port2) again, because port2 is truly defunct (is not even associated with the underlying message queue any more) |
| 20:17 | <Hixie> | i don't understand how we're supposed to tell the difference between a closed port and the remaining part of a cloned port |
| 20:18 | <atwilson> | Why not? closing a port is different from cloning it. |
| 20:18 | <atwilson> | They are completely different actions. |
| 20:19 | <Hixie> | All closing does is disentangle the ports |
| 20:19 | <Hixie> | Cloning does that too |
| 20:19 | <Hixie> | unless you want me to add some flag that remembers whether the port was closed or cloned |
| 20:19 | atwilson | is looking at the spec again while we talk |
| 20:20 | <atwilson> | Yeah, I'd say that's what I'm saying. We should differentiate between the two states, because they are different. |
| 20:20 | <atwilson> | A cloned port is no longer a task source |
| 20:20 | <atwilson> | a closed port is |
| 20:20 | <vvv> | What's up with my <input type=color alpha> proposal? |
| 20:21 | <Hixie> | atwilson: k, send mail |
| 20:21 | <Hixie> | vvv: it's number 296 out of 379 in the e-mail queue |
| 20:21 | <vvv> | I see |
| 20:23 | <annevk2> | "So, what you are going to create an element for every ARIA role now? That is crazy." -- you know, if you reverse things in that sentence that's pretty much how I thought about ARIA in the beginning (and still do regarding some parts) |
| 20:25 | <Philip`> | vvv: If you bug Hixie about it enough, he'll decide that it's less effort to respond immediately rather than keeping it in the queue :-) |
| 20:25 | <Philip`> | Sounds like the queue's quite short, though |
| 20:26 | <annevk2> | Philip`, do you already feel like it's next week? |
| 20:26 | <vvv> | Philip`: I'm delighted he at least said where it is in queue :) |
| 20:29 | <Philip`> | annevk2: Are you referring to something I said within the past few days? |
| 20:30 | <Philip`> | (I've spent most of that time period in bed or feeling like I'd rather be there, so I can't really remember what I may or may not have said) |
| 20:31 | <annevk2> | Philip`, you said you'd look into doing another charset survey on a larger dataset |
| 20:31 | <Philip`> | annevk2: Oh, okay |
| 20:31 | <annevk2> | I'm not feeling too well either (again)... :/ |
| 20:31 | <Philip`> | annevk2: "Next week" probably means Tuesday, in that case |
| 20:32 | <annevk2> | okidoki |
| 20:39 | <Hixie> | if you want to be at a lower point in the queue, you can file a bug for it |
| 20:39 | <Hixie> | of course, i'm not responding to bugs right now, i'm responding to e-mails |
| 20:39 | <Hixie> | so it's unclear which will get addressed first :-) |
| 20:40 | <Philip`> | Is there any penalty for sending an email and filing an identical bug? |
| 20:40 | <Hixie> | nope |
| 20:40 | <Philip`> | Sounds like a clear strategy, then |
| 20:43 | <Hixie> | vvv: (i haven't yet studied your e-mail, but off-hand it sounds similar to past proposals to which i responded that we would be better off waiting until we had experience with type=color before adding more features) |
| 20:43 | <Hixie> | ok, off to lunch |
| 20:43 | <Hixie> | bbl |
| 20:44 | <TabAtkins> | holy crap I need to go to lunch too. Nearly 3 hours overdue because I was so wrapped up in coding. >_< |
| 20:45 | <TabAtkins> | Thank Resig that jQuery makes it as easy as it is to create a responsive ajax-driven UI. |
| 23:37 | gsnedders | ponders buying web5.nu |
